Abstract
The end result of the laboratory process is the production of information. The quality of what is produced depends not only on quality factors such as the reliability of the information but also its turnaround time, accessibility and production cost.
Laboratory automation first started at the instrument level with the automation of sample analysis and well before microcomputers became part of the solution. Next followed Laboratory Information Management Systems (LIMS) to manage in particular the volume of instrument generated results.
The initial LIMS were based on minicomputers to which instrument connection capabilities were subsequently added, expanding in turn their information management powers. The first instruments were electro-mechanically oriented, then microcomputers replaced multiple electronic components, the current evolution being largely based on data processing managed on PC based workstations allowing Graphical User Interface (GUI) for both analytical process management and data handling.
With the standardisation of the GUI and the decreasing costs of corresponding off the shelf products, instrument workstations are now able to manage all workbench activities and for all disciplines. The opportunity however, for continued major productivity gains at this level appears to be limited as only incremental improvements are now forthcoming in this area. As a consequence laboratories and manufacturers are focusing on other aspects of sample management, particularly pre and post analytical activities, for the next quantum leap in productivity gains.
Total Laboratory Automation is then the logical extension of workbench automation. However, as the number of parameters to be controlled in a given configuration is expanding rapidly, key requirements for the proper management of the multiple devices potentially involved are the quality of the user interface and data accessing capabilities. The user interface is not the only interface to be taken into consideration since a new breed of workstations evolving to manage such configurations — A Workcell Manager— are surrounded by other interfaces like host computer (often LIMS) and the multiple device protocols as depicted in Figure 2. Standardisation of these multiple interfaces is one of the key issues still to be addressed in order to decrease the cost of total laboratory automation solutions and facilitate a modular automation approach. The Workcell Manager could however, be considered as the shell of the entire configuration, the user interface being the most complex part to be designed.
The Clinical Laboratory - An Information Business
Instrument Workstation/Workcell Architecture
To fullfill some of the Workcell Manager requirements, a large part of instrument workstation design concepts can be reused like the “dashboard” concept, giving at a glance the general status of the entire configuration. The hypertext approach, to navigate from the higher levels of information down to required details is also needed. Only relevant information should be brought to the attention of the operator. The user interface look and feel is also of paramount importance to zoom on the individual components like sample feeder, centrifuge unit, decapper, aliquot production unit, bar code production/labelling unit, sample conveyor as well as the various instruments for which both analytical process and data handling interfaces need to be considered. In addition, the Workcell Manager should also be able to receive the LIMS and/or host computer user interfaces. Again, in order to be in a position to implement a centralized user interface for such a large panel of components, the availability of standards (protocols and/or API, style guide) is required. If not, only engineering type solutions will be offered with corresponding costs and extended implementation/delivery time frames.
One of the main difficulties to build an efficient user interface satisfying Workcell Manager requirements, is the available screen space for presentation which is always limited and the response time taking into account the considerable level of data to be managed (particularly for large configurations). Also, the ease of operation is important when targeting multiple types of operators and multiple specialities of the laboratory which can both be automated and centralized. A positive outcome is that the same technics can be applied to Point Of Care instrumentation (standards, remote access for both analytical process and data handling) which can further improve the efficiency and adaptability of the proposed products for both users and manufacturers. The efficiency of any configuration is largely dependent on the reliability of the different sub-systems for which preventive actions and troubleshooting should be computer aided. Indeed, the already mentioned quality factors (reliability, turnaround time, accessibility and production cost) can all be impacted by the breakdown or misfunctioning of a component. For example, instrument failure can affect the reliability of information and its turnaround time; data processing failure can affect accessibility and again turnaround time, both impacting the production cost. On-line maintenance documentation including precise operating instructions, drawings and even motion video can greatly assist to reduce the time to repair. In addition, remote access to expert maintenance centers can further improve the time to repair. Also automatic processes can be triggered to occur and in the last resort operator intervention requested, both cases requiring a proper user interface. This subject area also implies a standardisation process, leaving room for proprietary data, in order to optimize development and operation costs. In addition, the way the large volume of potential errors is handled can significantly impact the general throughput and production of the whole system. Appropriate dashboard warnings integrated with efficient hypertext type navigational aids are key elements that must be taken into consideration.
In summary, just about all required components are readily available on the market in order to build a total laboratory automation solution. However, both the lack of standards and the lack of remote control capabilities makes each configuration unique and prevents the user interface from being really efficient and consistent. Without significant improvements in these two aspects, the risk is to see again the situation that we faced during the last decade for instrument workstations and in particular instrument connections where numerous proprietary solutions often offered only partial solutions to the laboratory's needs.